改进的免疫克隆选择算法及其在多峰值寻优中的应用

为了解决Castro克隆选择算法中存在的种群规模需根据经验确定、多峰搜索能力弱、训练时问长的问题，提出了一种新的免疫克隆选择算法，该算法基于实数编码和自适应变焦变异方法，能够动态确定种群大小，具有很强的全局和局部搜索能力，可以搜索到全局最优点和尽可能多的局部极值点。仿真实验结果表明，改进的算法平均运行时间和平均找到的峰值点个数都明显优于Castro克隆选择算法，多峰值函数的优化效果得到了显著改善。

Introduction to the revised immune colonial selection algorithm and its application in optimization under multi-peak model

A revised immune colonial selection algorithm is put forward to resolve the problems arising from the Castro colonial selection algorithm, such as experience dependency in the confirmation of population size, weak multi -peak search capability, long training period, etc. This new algorithm, based on real coding and self-adapting zoom and variation, is able to confirm population size under dynamic mode and find global optimum and local ex- tremes as many as possible with a strong global and local search capability. The result of the simulated experiment shows that the operation time used and the number of extremes found on average by the revised algorithm are more outstanding than those of Castro colonial selection algorithm, the optimization of multi - peak function is improved.